On 11/03/2014 16:40, Simon Urbanek wrote:
On Mar 11, 2014, at 10:35 AM, Stephen B. Cox <[email protected]> wrote:
Anyone had any experience running fairly intensive analysis on a new
MacPro? I am looking to upgrade my desktop, and 80% of my time is spent in
RStudio/Latex/Sweave... working primarily with microbiome analysis (large
datasets). Been considering a new MacPro, but I am a little hesitant
about; a) moving my desktop to Mac,
'large datasets' are in the eye of the beholder: you would need to
quantify that.
That is typically a big plus - especially if you use Windows. It is in fact
probably the single major reason to pick a MacPro today, although I would
probably rather get an iMac in that case.
and b) whether the MacPro performance will be worth the cost (it almost seems
geared more towards graphics than anything else).
I don’t have any hands-on experience with the new MacPro but its specs are
underwhelming. It is an experiment - if you can leverage the GPUs for
computing, then it may be worth it, but it’s still quite hard to do so. With
AMD you’e essentially stuck with OpenCL and other than core support so you can
write your own kernels, there is very little else in R to leverage that. Today,
you’re much better off getting a server/workstation which you can load with RAM
and more cores for computing for the same price (running Linux, obviously, you
really don’t want to do computing on Windows with R) - and use your
desktop/laptop just to access its computing power.
For some background - I have worked on Macs for years, but moved my main work
desktop to Windows about 2 years ago. I also do quite a bit of work in QIIME -
which can be done on the mac (not the PC) and is both RAM and CPU intensive...
so, I can benefit from multiple cores, large RAM, etc. My 2011 MacBook Pro
seems extremely sluggish at this point when running basic tasks (probably need
to do a fresh OS install),
If you encounter sluggishness in OS X is pretty much always a disk issue. Wipe
the disk or even better put in a SSD - it’s more than worth it - a whole
different world.
Or a 'fusion' drive in an iMac, which gives you enough SSD advantage
unless you really use repeatedly a lot of disc space (and works well for
me). The MacPro's I/O benchmarks are impressive, but you would need to
be able to generate data at those speeds to make use of them.
Cheers,
Simon
but the Windows machine has
never slowed down. This has added to some of my hesitation.
Anyone have opinions/experience using R on the new MacPro?
On Mon, Mar 10, 2014 at 1:05 PM, Simon Urbanek
<[email protected]>wrote:
On Mar 10, 2014, at 12:43 PM, Nick <[email protected]> wrote:
Good afternoon, I am looking at buying my first Mac and thought i'd ask
for advice for what I should get. I have it down to the two models below
(but am open to realistic suggestions).
I will primarily be using R for machine learning packages, and the data
sets are very large. If any other specs are needed let me know.
"data sets are very large' - well, the machines listed below are certainly
not suitable to run anything on large data ;) so you may want to quantify
what you mean here. You want as much RAM as possible for large data since
that is the single item that will cause huge drop-off in performance when
exhausted and R certainly can take quite a bit of memory if this is really
your only machine to run computing on. Note that in modern Apple laptops
you cannot add more memory later, so this is rather important factor.
Given a choice of the two MacBook Air is not a computing machine - it's
optimized for power consumption, not speed, so the only reason to go for it
is if you're looking for a light notebook and don't care about the
computing speed as much.
Cheers,
Simon
Thanks in advance.
13-inch MacBook Air ($1,349)
1.7GHz Dual-Core Intel Core i7, Turbo Boost up to 3.3GHz
8GB 1600MHz LPDDR3 SDRAM
128GB Flash Storage
13-inch MacBook Pro with Retina ($1,399.00)
2.4GHz Dual-core Intel Core i5, Turbo Boost up to 2.9GHz
8GB 1600MHz DDR3L SDRAM
128GB PCIe-based Flash Storage
[[alternative HTML version deleted]]
_______________________________________________
R-SIG-Mac mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-mac
_______________________________________________
R-SIG-Mac mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-mac
[[alternative HTML version deleted]]
_______________________________________________
R-SIG-Mac mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-mac
_______________________________________________
R-SIG-Mac mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-mac
--
Brian D. Ripley, [email protected]
Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/
University of Oxford, Tel: +44 1865 272861 (self)
1 South Parks Road, +44 1865 272866 (PA)
Oxford OX1 3TG, UK Fax: +44 1865 272595
_______________________________________________
R-SIG-Mac mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-mac
